#✨│ai-help
timid ruin
ich finde das nicht irgendwie=
habs sorry man muss auf "Start"
ist das so richtig?
Muss ich dann aber bei den ganzen anderen Modulen von vorne beginnen?
also das alles neu ausfüllen wahrscheinlich oder?
pastel oak
Muss ich dann aber bei den ganzen anderen Modulen von vorne beginnen?
Da du neues datenset hast musst du wieder pre processen, feature extraction etc. machen
pastel oak
also das alles neu ausfüllen wahrscheinlich oder?
Ja, mit dem gleichen voice model namen dann
alpine oyster
Merhabalar benim bir arkadaşımın ses modelini çıkatmak istiyorum bana yardımcı olabilecek varmı?
timid ruin
Okay also oben bei Name einfach den gleichen Namen. Alles einstellen wie gewohnt und laden lassen nur mit den neuen Datein die ich nutzen möchte?
alpine oyster
Ja, mit dem gleichen voice model namen dann
Selam yazdığım ile ilgili yardımcı olabilirmisin
pastel oak
Okay also oben bei Name einfach den gleichen Namen. Alles einstellen wie gewohnt...
Genau und dann trainieren und es setzt fort an der stelle an der du zuletzt aufgehört hast
alpine oyster
Selam yazdığım ile ilgili yardımcı olabilirmisin
?
alpine oyster
Teşekkürler
timid ruin
Genau und dann trainieren und es setzt fort an der stelle an der du zuletzt aufg...
ich glaube ich habe es verstanden dankeschön :)
Noch eine Frage ich hatte den ersten Versuch mit nur 1 Minuten Audio gemacht & danach jetzt mit 4 Minunten. Aber die 1 Minuten Audio hört sich deutlich deutlich besser an als die zweite? Woran kann das liegen? Bei der zweiten habe ich auch noch drauf geschaut das keine Musik nichts ist, nur Text.
pastel oak
ich glaube ich habe es verstanden dankeschön :)
Noch eine Frage ich hatte den e...
Wie meinst du die zweite hört sich besser an? Das Datenset, oder die Stimme die danach rausgekommen ist?
timid ruin
Wie meinst du die zweite hört sich besser an? Das Datenset, oder die Stimme die ...
die Stimme vom ersten Model mit 1 Minute Audio hört sich besser an als die Stimme mit 4 Minuten Audio
1 Versuch: 1 Minuten Audio
2 Versuch: 4 Minuten Audio mit kaum Hintergrundgeräuschen
1 Versuch hört sich viel besser an als der zweite?=
pastel oak
die Stimme vom ersten Model mit 1 Minute Audio hört sich besser an als die Stimm...
kommt drauf an wie viele steps und epochen du schon durch hast. die 4 minuten dauern länger zu trainieren würde ich denken
kann auch an der batch size liegen die man evtl. anpassen muss.. kann auch nurn zufall sein
timid ruin
kommt drauf an wie viele steps und epochen du schon durch hast. die 4 minuten da...
Wie kann ich die Batch Size anpassen? Und wie sollte ich jetzt am besten weiter machen?
Tut mir leid bin echt neu :/
das hier?
merry eagle
is it necessary to install apolio when i open this page. because i already done it yesterday?
pastel oak
Wie kann ich die Batch Size anpassen? Und wie sollte ich jetzt am besten weiter ...
Ich habe persönlich noch nie eine stimme mit einem veränderten datenset trainiert deswegen weiß ich das nicht :x
Ja das ist das batch size, allerdings weiß ich nicht wie es sich mit einem geänderten datenset aufs ganze auswirkt
was für eine gpu hast du?
timid ruin
was für eine gpu hast du?
RTX 2060
pastel oak
is it necessary to install apolio when i open this page. because i already done ...
I believe you have to run these things everytime because its not cached. You can try to skip it and try if it works, if not then yes you do have to do it again
timid ruin
Ich habe persönlich noch nie eine stimme mit einem veränderten datenset trainier...
ohh okay
timid ruin
is it necessary to install apolio when i open this page. because i already done ...
is that still working?
one guy in the yt tutorial siad its not working anymore
merry eagle
is that still working?
you can see in image i'm installing this time. next time i don't install and check
merry eagle
one guy in the yt tutorial siad its not working anymore
nope this one is working for me perfectly
timid ruin
so ist alles richtig oder?
pastel oak
one guy in the yt tutorial siad its not working anymore
Dont trust people of youtube videos pls
timid ruin
hm ja ich glaub 4 minuten batch size 8 sollte passen
ich versuche es jetzt einfach mal
pastel oak
so ist alles richtig oder?
mach die anführungszeichen zur sicherheit nochmal weg
timid ruin
mach die anführungszeichen zur sicherheit nochmal weg
ups hast recht
timid ruin
Dont trust people of youtube videos pls
Also der hatte auch Recht. Ich habe es probiert und dauernd kam ein Bot alert
timid ruin
nope this one is working for me perfectly
ohh thats good! No bot alert?
merry eagle
ohh thats good! No bot alert?
No bot alert
rare gobletBOT
Ayo? @merry eagle level 3 !!! 
timid ruin
No bot alert
Ohh okay cool! Which tutorial are you using?
rare gobletBOT
Ayo? @timid ruin level 4 !!!
pastel oak
merry eagle
Ohh okay cool! Which tutorial are you using?
you don't need any tutorial in this bot. it's pretty easy.
timid ruin
you don't need any tutorial in this bot. it's pretty easy.
ohh okay :D ty
timid ruin
https://docs.aihub.wtf/rvc/cloud/applio-colab/
Tutorial? :
pastel oak
Tutorial? :
ja
timid ruin
danke
merry eagle
Tutorial? :
it's the easy one
timid ruin
ja
meine Stimme klingt so arsch. Wie kann ich das besser machen? Neue Audio?
mein dataset ist auch echt gut eigentlich :/
leaden yacht
f0fail-25-T:\Applio\ApplioV3.1.1\logs\Glamrock-Freddy/1_16k_wavs/95_0.wav-Failed to load audio: [WinError 1455] The paging file is too small for this operation to complete
This happened with a lot of files, what do I do about this?
wispy burrow
f0fail-25-T:\Applio\ApplioV3.1.1\logs\Glamrock-Freddy/1_16k_wavs/95_0.wav-Failed...
Did you split your Dataset yourself?
wispy burrow
No I got it straight from the game files
Were some of them shorter than 1 sec?
leaden yacht
nah
defi not
wait imma recheck
but this also happened with another dataset where every file was 5-30 seconds
leaden yacht
Were some of them shorter than 1 sec?
are files that show more than 1 sec good tho?
or should I js paste them all into one big file?
nimble flax
what do i put here, i have a gtx 1060 6gb
rare gobletBOT
Ayo? @nimble flax level 3 !!!
sonic harbor
Wait someone tell me how do i make an ai cover pls 😭
thin stump
Wait someone tell me how do i make an ai cover pls 😭
Check the channel pins before asking for help. It's there for a reason
thin stump
wispy burrow
are files that show more than 1 sec good tho?
Well, you could create one big file.
I had lots of files files that were really short.
I got the same error then.
I put them in Audacity and created 3 to 5 clips out of them.
And then it worked.
I hadn't thought of creating 1 file.
Would have saved me quite some time...
leaden yacht
Someone already helped me and the errors are fixed now, thanks though
It doesn't matter how big the file is, even 1 second is good. Its only bad if it is really to close to 0 seconds
azure patio
what would be the most accurate rvc
nimble flax
what do i put here, i have a gtx 1060 6gb
??
limpid thistle
i installed CServer IO the ai was working
but when i restarted my pc
the http doesnt wanna launch
anyone can help
wispy burrow
??
Maybe ask in #🔍│help-w-okada .
I know it isn't W-Okada.
It's realtime RVC I think
But that is the channel for Realtime Voice Conversion.
wispy burrow
anyone can help
That's W-Okada.
Try asking in #🔍│help-w-okada
low shard
cause for me, public url isn't showing anymore
show a screenshot
frank shuttle
give me a sec to redo everything
rare gobletBOT
Ayo? @frank shuttle level 4 !!!
frank shuttle
show a screenshot
Usually it shows the public link very fast after it shows the local link
low shard
lemme check rq
frank shuttle
So yeah it is probably broken rn
I will just use Applio until there is a fix
nevermind
vale osprey
you get it to work?
nocturne mural
gradio share down
frank shuttle
you get it to work?
no
low shard
gradio share down
wait what
silk jasper
Guys where is the turkish sounds
low shard
So does that mean if I try local it also won't work?
it does work locally
frank shuttle
Ok I will try that then
nocturne mural
So does that mean if I try local it also won't work?
in local of course it will work since the localhost is there, but in colab you need a separate link since you cannot access the colab's localhost just like that.
frank shuttle
I understand about two words in that sentence
not that your not clear with what you are saying
I am just an idiot
nocturne mural
i speak spanish xd
in other words, in local it will work correctly but in the colab it is accessed in another way, and that other way is through the "share link".
light bolt
how to fix WinError 126
brittle wing
uh how do i set the input to the voice changer for fortnite
anyone know why my shits kinda laggy?
i fucked around with the chunks
but they all kinda miss out words/lag over words
so it cuts part of the sentence
dusty lava
is gradio down again?
south kiln
help please
i am new
i have an issue with a thing
my problem is like the previous ones
rare gobletBOT
Ayo? @south kiln level 1 !!!
storm rose
i went to create an AI cover on IlariaRVC on GoogleColab but it doesn't show any public link to create it, does anyone know why?
rare gobletBOT
Ayo? @storm rose level 2 !!!
storm rose
or like, anywhere else i could do this?
wind oar
@odd shale Can someone help me with training ur own model?
more like help me since I am stuck at this step to import my training folder to start processing data
rare gobletBOT
Ayo? @wind oar level 1 !!!
nimble flax
the voice is laggy on discord
im having the same issue
storm rose
is gradio down again?
i think so, no public link appears when i go try to do an ai cover
violet heron
the voice is laggy on discord
Is it laggy on the voice changer itself @nimble flax @south kiln
nimble flax
no it isnt
violet heron
no it isnt
Are you using the VB Audio cable
nimble flax
yes for discord
violet heron
Try using VAC Lite instead
verbal vapor
how do i do this
?
azure marshBOT
RVC Guides (How to Make AI Cover)
verbal vapor
i saw that too but i dont think u can u just gotta wait till gradio back up
dang alr
junior halo
Is there any other platform I can test my voice model on? I've run out of compute units and I don't feel like buying any.
silver basin
Same issue with Ilaria-RVC, is there another working Colab?
violet heron
Is there any other platform I can test my voice model on? I've run out of comput...
HuggingFace Spaces or locally on your computer
cloud sandal
Me
light bolt
help
rare gobletBOT
Ayo? @light bolt level 1 !!!
light bolt
how to fix
oserror: [winerror 126] the specified module could not be found
merry tundra
someone can help me in vc for RVC installation?
violet heron
someone can help me in vc for RVC installation?
Download zip depending on your GPU
Unzip and run the .py file in python
Or the bat file for go-web if it’s there
merry tundra
wait i fixed
but now they say i need a file .PTH for the voice model
but i have only the rar link
@violet heron
merry tundra
wait, i did it but it doesn't show up
rare gobletBOT
Ayo? @merry tundra level 2 !!!
violet heron
wait, i did it but it doesn't show up
What files are in there
late swallow
-colab
strong fulcrum
What's the most expressive TTS? (with ai like tortoisetts or without)
Or what's the best way to make tts the most expressive like with rvc?
violet heron
and i put them in the asset folder
You put the .pth in the weights folder
And the .index in logs folder
tall depot
What's the most expressive TTS? (with ai like tortoisetts or without)
style bert vits2
violet heron
i am using voice.ai, the voice changer ain't working no more
Ewwwwwwwwwwww
Don’t use voice ai, it’s paid and proprietary
Use w-Okada instead for real time or RVC for covers
They are wayyyy more customizable, completely free and open source, and give you way more freedom
junior halo
I am so confused. I am trying to complete this voice model of Ordan Karris from Warframe. The first sample I trained on a 9 minute audio clip of him speaking from a YT video. 350 epochs, 8 steps. The second I used ripped in-game .wav files on 500 epochs, 10 steps. For some reason, the second one sounds worse.
stable crest
Localhost not here help
mortal bolt
why do my voices sound robotic?
dawn nebula
Localhost not here help
what
dawn nebula
Localhost not here help
pls send image of ur console
junior halo
dawn nebula
why do my voices sound robotic?
bad model, or wrong settings
mortal bolt
Yeah I made a voice model how do I make it better?
stable crest
pls send image of ur console
rare gobletBOT
Ayo? @stable crest level 2 !!!
dawn nebula
what gpu do u have
stable crest
what gpu do u have
Intel HD Graphics 400
odd shale
They are wayyyy more customizable, completely free and open source, and give you...
Hello Jimbovill
Tangoville
stable crest
i think just use colab
I rather use this https://rentry.co/TrainingVoiceModels
-local
azure marshBOT
thin stump
I am so confused. I am trying to complete this voice model of Ordan Karris from ...
You said you used a 10 batch size?
always stick to 2's so 4, 8, 12, 16
dawn nebula
I rather use this https://rentry.co/TrainingVoiceModels
u cant, since u dont have a nvidia gpu
fossil rover
-colab
gusty kestrelBOT
@fossil rover, I have found 1 results that match your search!
Requested by cheffy.
@fossil rover, I have found 1 results that match your search!
Requested by cheffy.
@fossil rover, I have found 15 results that match your search!
Requested by cheffy.
@fossil rover, I have found 1 results that match your search!
Requested by cheffy.
@fossil rover, I have found 5 results that match your search!
Requested by cheffy.
@fossil rover, I have found 1 results that match your search!
Requested by cheffy.
@fossil rover, I have found 1 results that match your search!
Requested by cheffy.
@fossil rover, I have found 2 results that match your search!
Requested by cheffy.
azure marshBOT
RVC Guides (How to Make AI Cover)
magic rivet
does anyone have a link to a colab that can do batch AI covers?
I can see the option on local versions, but my computer is too crappy to run on local
warm depot
-help
warm depot
-rvc
warm depot
how do i get help here
thin stump
how do i get help here
In the most polite way possible, you need to state your problem
dreamy cliff
-colab
magic rivet
Hello hello! Does anyone know if there is a colab RVC that supports batch outputs?
violet sky
yo if i keep re-running the vocals through rvc after every output does it work that way?
azure marshBOT
carmine hearth
By default, hugginface space uses CPU. If you want to use the GPU, you need to pay extra.
https://huggingface.co/docs/hub/spaces-gpus
magic rivet
does anyone know what Scheme missing. means
brave garnetBOT
⠀
Google Colabs 
⠀
AICoverGen-WebUI
Useful for making quick covers, by Hina.
AICoverGen-NoWebUI
Useful for making covers, doesn't include a UI, by Ardha, by Eddy, Hina and Gdr.
RVC Disconnected
To train new voice models, by Kit Lemonfoot.
EasyGUI
The OG interface, by Rejects.
⠀
Made with love by Ilaria 💖
brave garnetBOT
⠀
Google Colabs
⠀
AICoverGen-WebUI
Useful for making quick covers, by Hina.
AICoverGen-NoWebUI
Useful for making covers, doesn't inclued a UI, by Ardha, by Eddy, Hina and Gdr.
RVC Disconnected
To train new voice models, by Kit Lemonfoot.
EasyGUI
The OG interface, by Rejects.
⠀
Made with love by Ilaria 💖
⠀
Google Colabs
⠀
AICoverGen-WebUI
Useful for making quick covers, by Hina.
AICoverGen-NoWebUI
Useful for making covers, doesn't inclued a UI, by Ardha, by Eddy, Hina and Gdr.
RVC Disconnected
To train new voice models, by Kit Lemonfoot.
EasyGUI
The OG interface, by Rejects.
⠀
Made with love by Ilaria 💖
azure marshBOT
RVC Guides (How to Make AI Cover)
placid fossil
-realtime
azure marshBOT
-realtime
This interaction has expired, use the command
/guides realtimeif you wish to see it again.
zenith comet
help rvc
azure marshBOT
🎙️ Realtime Voice Changer (Local)
fluid matrix
help rvc
bro please be a bit more specific, what exactly do you need help with. what are you using? any error codes?
zenith comet
bro please be a bit more specific, what exactly do you need help with. what are ...
yes finally thank you i downloaded the rvc gui and it gives an error code like this.
"
C:\Users\ASUS\Desktop\RVC-GUI-Windows-pkg>python.exe rvcgui.py --pycmd python.exe
Backtrace (last call):
File "C:\Users\ASUS\Desktop\RVC-GUI-Windows-pkg\rvcgui.py", line 4, in <module>
import soundfile as sf
ModuleNotFoundError: There is no module named 'soundfile'
C:\Users\ASUS\Desktop\RVC-GUI-Windows-pkg>pause
Press any key to continue . . .
"
Translated with DeepL.com (free version)
rare gobletBOT
Ayo? @zenith comet level 1 !!!
fluid matrix
yes finally thank you i downloaded the rvc gui and it gives an error code like t...
delete rvc gui, super outdated
what do you want to do?
train a model or infer something? and what are your specs
zenith comet
delete rvc gui, super outdated
I'm just gonna make a noise and my specs are RTX 3060 TI i7 12900k
fluid matrix
make a noise?
you mean a cover of something?
infering
anyways, follow the guides on the docs instead
your specs are good enough for a local solution so use one of those
zenith comet
make a noise?
I'm talking to you through the deepl translation app, it might have translated wrong, sorry, I apologize.
fluid matrix
no problem
proud elbow
I'm just gonna make a noise and my specs are RTX 3060 TI i7 12900k
it's good already for local inference & training, I'd recommend either mainline RVC or applio
fluid matrix
yep
zenith comet
it's good already for local inference & training, I'd recommend either mainline ...
rvc gui always gives an error when trying to start.
proud elbow
rvc gui always gives an error when trying to start.
nah I mean this RVC https://docs.aihub.wtf/rvc/local/mainline/
you should try it
last olive
@finite galleon bro when i am using apolio rvc the public url changes in every 10-15 min. and then url changes i have to restart my whole work again.
finite galleon
<@818050831034613771> bro when i am using apolio rvc the public url changes in e...
U on google colab?
last olive
U on google colab?
yes
finite galleon
yes
Probably something related to this thats causing the issue #📰│dev-updates message
last olive
Probably something related to this thats causing the issue https://discord.com/c...
can you suggest me any alternative
finite galleon
can you suggest me any alternative
Ig you can use applio no ui https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio_NoUI.ipynb
Its harder to use but ig your best bet until the gradio thing is fixed
last olive
i am entering this link
rare gobletBOT
Ayo? @last olive level 1 !!!
last olive
bro that's not the issue. The issue is the link is changing again and again and i have to restart colab the again and again
can't convert
sometimes can't refersh and get internet connection error
uploading take too much long time
hot lagoon
Which use the BS Roformer latest one ?
crimson badge
Whenever I download a model on the Ilara RVC thingy, no matter how many times i press refresh, the model wont show up
I made sure the link was public
hot lagoon
Wich is that on uvr or?
last olive
so which mark are you talking about
finite galleon
Wich is that on uvr or?
That is the zfturbo finetuned version of viperx bs roformer and is not available on uvr
On uvr just use the viperx bs roformer model
azure marshBOT
RVC Guides (How to Make AI Cover)
turbid meteor
so, rvc crash when i start it. i think its because i selected wrong input and output device
how do i know if im using the right input/output? ive tried them all (windows direct sound, wasapi, wdm-ks) and didnt work
my mic is cm383-080864
last lava
That is the zfturbo finetuned version of viperx bs roformer and is not available...
bro how to pretrain and what is the advantage doing this before training the voice?
rare gobletBOT
Ayo? @brittle wing level 1 !!!
fiery raptor
Hi
still looking for help
Maybe saying everyone's favorite buzzword will work
Ilaria
fiery raptor
Hi I still need help btw
heady skiff
I made a model on weights gg. and it sounds great, but I don't want to use it anymore on this site and want to use this model elsewhere. I downloaded it from the site. Where and what can I do as an alternative?
rare gobletBOT
Ayo? @heady skiff level 2 !!!
thin stump
I made a model on weights gg. and it sounds great, but I don't want to use it an...
The weights devs say they were looking into removing the index file from models created with weights.gg. You only get the pth file unfortunately
frank shuttle
I clicked run_applio, command opened, nothing happened
I am still having this problem
pls help
thin stump
Unless it does give you index then I'm wrong
fiery raptor
why did this happen
Im still having this problem
heady skiff
what is an index file? Sorry haha
I do see a pth file in the folder of my downloaded model
thin stump
what is an index file? Sorry haha
It's the accent for the model so you would need that for real time voice changer, or any huggingface colab
heady skiff
It's the accent for the model so you would need that for real time voice changer...
do people use weights gg for ai voices? Sorry I'm super new and I guess the first thing I jumped into immediately was weights gg, which was great and I've made great voice models, but the queue kinda sucks. I'm looking for a more immediate result. (using it for voice acting, not covers.)
spring gate
I don't know why I keep getting this
heady skiff
I guess overall I'm using AI voices for voice acting and waiting 19 minutes for an audio less than a minute is sorta weird to me? Is that average for people?
frank shuttle
it all depends on what you use
like there are times that I can get a 3 minute audio done in 15 seconds
heady skiff
I'm using weights gg 
Just now learning that they don't give you the appropriate files once you download your voice model
heady skiff
like there are times that I can get a 3 minute audio done in 15 seconds
that's exactly what I'm looking for
frank shuttle
damn
frank shuttle
that's exactly what I'm looking for
The best ways to do it are in here
both local and online
thin stump
do people use weights gg for ai voices? Sorry I'm super new and I guess the firs...
Ai hub has a partnership with weights.gg because it's free and the devs actually communicate unlike kits or voicify. And I recommend creating voice models on RVC disconnected or local (GPU). He gave you the link above^^
frank shuttle
Ai hub has a partnership with weights.gg because it's free and the devs actually...
Do you know why this problem happens when I try to run applio locally? #✨│ai-help message
heady skiff
Ai hub has a partnership with weights.gg because it's free and the devs actually...
Thanks a lot! Is this a certain program or is this just different ways to do it with different methods?
thin stump
Do you know why this problem happens when I try to run applio locally? https://d...
Nope, I don't use Applio
fiery raptor
I dont use applio either
thats why I didnt use this tip from Light
cuz it said "Start applio" and I dont use that
so
frank shuttle
I usually use Ilaria but wanted to try something local
does anyone know if the gradio public links are fixed now?
frank shuttle
damn
fiery raptor
fun fact
rare gobletBOT
Ayo? @fiery raptor level 5 !!!
fiery raptor
Ive been here for an hour and a half
frank shuttle
wdym by that
fiery raptor
Im still waiting for help
heady skiff
https://docs.aihub.wtf/
wait, what program do you use to make your audios? sorry aa
thin stump
Thanks a lot! Is this a certain program or is this just different ways to do it ...
It's a better method than weights because you have control over your dataset. The guide has it all on dataset cleaning. Weights models are somewhat airy because they just expect you to toss in a audio sample and pass a requirement check
frank shuttle
wait, what program do you use to make your audios? sorry aa
Well I usually use Ilaria, but it has kinda been down since yesterday so I am trying other things
heady skiff
Well I usually use Ilaria, but it has kinda been down since yesterday so I am tr...
is it free? the way you can make stuff instantly is what I'm looking for
frank shuttle
Yes
rare gobletBOT
Ayo? @frank shuttle level 5 !!!
south kiln
i have a problem who could help me?
burnt zealot
can someone send me the file link to voice train locally with rvc
sharp vale
im trying to download rvc from google colab and JSONDecodeError: Expecting value: line 1 column 1 (char 0) is still showing up
can someone help me?
heady skiff
Think I'm gonna try using llaria as well but I'm unsure if I can use my weights gg model 
frank shuttle
I am not sure if it is back up tho
Cause yesterday the website it uses went down
kinda
south kiln
just him
frank shuttle
hello
south kiln
hi
i have an issue
i was trying the voice changer on discord
i got the cable or mth like that
frank shuttle
wdym the voice changer thing
south kiln
yeah the one that u use to prank friends on discord
south kiln
and yet i cannot anymore due to an issue linked to the voice cable
frank shuttle
Oh
south kiln
i got the voice setting of discord set on the cable
the input is set on cable not microphone
output on speakers
but by doying the record test it doesnt hear anything
how
frank shuttle
What is interesting
is that I am having that same problem with one of my mics
but when I switch the the terrible mic built into my headset, it works fine
south kiln
ah so i just need to change mic
frank shuttle
Did you check YT for solutions?
south kiln
yeah and i onyl found DUckus
frank shuttle
Damn
south kiln
Duckus vid on this
but he just tells your what to do
and yesterday it worked
after i plugged in the pc back it doesnt work anymore
crazy stuff
frank shuttle
that does not make any sense
south kiln
ikr
frank shuttle
there has to be a fix somewhere
south kiln
yeah
frank shuttle
maybe try reaching out to Discord support
south kiln
yes but i have no time fot searching ;_;
yk it was somethign that if it works it works if not amen
ty anyway
burnt zealot
maybe try reaching out to Discord support
do u have the rvc local download link?
low shard
do u have the rvc local download link?
check our docs https://docs.aihub.wtf
long gazelle
how do i download rvc gui?
rocky maple
is there any ai cover gen that has a setting for vocal only inputs besides weights.gg?
proper shale
how do i download rvc gui?
RVC GUI is outdated but u can use other stuff
Ofc there's always realtime:
proper shale
is there any ai cover gen that has a setting for vocal only inputs besides weigh...
any RVC fork does that
long gazelle
https://docs.aihub.wtf/
RVC GUI is outdated but u can use other stuff
Ofc ther...
why gui outdated?
rocky maple
any RVC fork does that
everytime i put in a vocal to an rvc fork it only does like half of the vocals
worst way of explaining it mb
proper shale
why gui outdated?
crepe only
long gazelle
ah
rare gobletBOT
Ayo? @long gazelle level 1 !!!
rocky maple
but like a gen that makes the cover regardless of input audio background noise ig?
long gazelle
why would realtime be better?
proper shale
everytime i put in a vocal to an rvc fork it only does like half of the vocals
keep in mind you don't need background vocals
proper shale
why would realtime be better?
thats an option for u
if u want that
it's not "better", just different use case
brittle wing
I think the program does not detect my microphone, how can I solve the problem?
rare gobletBOT
Ayo? @brittle wing level 1 !!!
pastel oak
I think the program does not detect my microphone, how can I solve the problem?
send screenshot of voice changer, send screenshot of the command window (black background, white text) and show discord voice settings
brave garnetBOT
⠀
Google Colabs
⠀
AICoverGen-WebUI
Useful for making quick covers, by Hina.
AICoverGen-NoWebUI
Useful for making covers, doesn't inclued a UI, by Ardha, by Eddy, Hina and Gdr.
RVC Disconnected
To train new voice models, by Kit Lemonfoot.
EasyGUI
The OG interface, by Rejects.
⠀
Made with love by Ilaria 💖
rotund cairn
i have a ? why in the videos i see with okada voicechanger theres is so smooth mine is ´good but not as good why i have decent specs rtx 2080 ryzen 7 3700x 32 gb ram 1 tb ssd
azure marshBOT
RVC Guides (How to Make AI Cover)
native wedge
how to I make the voice not shit
rare gobletBOT
Ayo? @native wedge level 1 !!!
native wedge
patent trellis
lusty knoll
-colab
quick jungleBOT
<@1158132286169022555>
Watchu want? :3
bitter tiger
@quick jungle how to setup the voice changer
quick jungleBOT
<@1158132286169022555> how to setup the voice changer
Watchu want? :3
bitter tiger
@mellow token how to setup the voice changer
@fluid matrix how to setup the voice changer
fluid matrix
pin everyone buddy im sure someone will answer
follow this guide
dont forget to install vb cable too
make sure your specs are good enough tho, you need a decent gpu
wheat plover
changing pitch_extraction_algorithm from RMVPE to Crepe does anything?
stable crest
u cant, since u dont have a nvidia gpu
How do you train a ai rvc model without an nvidia gpu?
violet heron
can u give me a link to a vid
-docs
violet heron
bout tofliP out cant get this stuff to work
!howtoask
median jetty
Trying to use this Lady Gaga voice from the voice models section here in the discord and easy gui is giving an error https://huggingface.co/Freddie-Dassin/Lady_Gaga/resolve/main/Lady Gaga.zip?download=true
azure marshBOT
Suggestions for @median jetty
rare gobletBOT
Ayo? @median jetty level 1 !!!
finite galleon
Do any of these allow to run locally with an AMD card? I have a RX 6800
Google colab is hosted on googles devices so you can run it on your laptop, phone, samsung smart fridge, whatever as long as it can open the internet
Oh wait i didnt read the locally part lmao
I am pretty sure most of the webuis dont support AMD
proud elbow
cant wait to see some DDoSer pull its puppet string
proud elbow
brittle wing
Google colab is hosted on googles devices so you can run it on your laptop, phon...
With a nokia phone 
proud elbow
Google colab is hosted on googles devices so you can run it on your laptop, phon...
nintendo 3ds
steel forge
What colab should I use to merge models? Applio is giving me a strange error...
trying ilaria next but idk if it has merging
azure marshBOT
median jetty
does use huggingface require an nvidia gpu?
steel forge
somewhat curious of this merge tool
https://cdn.discordapp.com/attachments/11592...
First reaction: AI Hub schitzo post
After reading: Holy shit I learned so much
that needs a pin or something
proud elbow
that needs a pin or something
it's codename's post that already pinned
steel forge
yeah its underneath the first pin in here
thats so much info, i love this
local mainline doesnt have 32k merging :(
thin stump
somewhat curious of this merge tool
https://cdn.discordapp.com/attachments/11592...
what i'm confused about is that my fm keeps rising while codename's example is always going down so I can't use fusion
well actually, it's different on the mangio fork (rvc disconnected) #✨│ai-help message
azure marshBOT
RVC Guides (How to Make AI Cover)
thin stump
yeah but it's good that I don't have to think of g/total as the end of all things. It's just the average and I know MJ was trying to drill that point in
proud elbow
what i'm confused about is that my fm keeps rising while codename's example is a...
I have no idea about fm, prob just look at g/total and mel
atomic bolt
start_https file wont load only start_http
sinful quiver
guys which one is better? rmvpe or crepe
finite galleon
guys which one is better? rmvpe or crepe
99.99% of the time rmvpe
Rmvpe is robust so it handles noisy audios pretty well but crepe could be better if the audio is really really high quality
white bough
Has anyone found a way to fix this out-of-range index problem on the RVC disconnected colab that happens with the Titan and the Ov2 32k pretrains? I don't mind if the fix is complicated, but I would like to try those ☺️
thin stump
Has anyone found a way to fix this out-of-range index problem on the RVC disconn...
Hi, for Titan 32k, you can use these two as the huggingface link: https://huggingface.co/blaise-tk/TITAN/resolve/main/models/medium/32k/pretrained/D-f032k-TITAN-Medium.pth?download=true
As for Ov2, I talked about it here #🔊│ai-development message You probably don't need it because Titan is much better soundwise
white bough
Hi, for Titan 32k, you can use these two as the huggingface link: https://huggin...
Thank you so much!! It is working ^^
dawn nebula
guys which one is better? rmvpe or crepe
rmvpe
fiery raptor
still nothing
pale gyro
hi, want to turkish singer YALIN sound
azure marshBOT
RVC Guides (How to Make AI Cover)
pale gyro
turkish
rare gobletBOT
Ayo? @pale gyro level 1 !!!
pale gyro
yes level 1
pallid patio
How can I download an RVC to make them without colab and etc?
low shard
How can I download an RVC to make them without colab and etc?
Check our docs https://docs.aihub.wtf
You will need a good gpu for training tho
pallid patio
Check our docs https://docs.aihub.wtf
I've checked it and downloaded an RVC, but it just don't open ;-;
low shard
I've checked it and downloaded an RVC, but it just don't open ;-;
Which rvc and what is the error
U got python installed?
pallid patio
Mangio, I've opened in the "web.bat", but it don't get me anywhere
pallid patio
U got python installed?
I didn't read about that anywhere. Can you explain me?
low shard
Mangio, I've opened in the "web.bat", but it don't get me anywhere
Also, it's better you use applio, btw can u show a screnshot of the error?
pallid patio
Also, it's better you use applio, btw can u show a screnshot of the error?
I'm trying it now, so I unstalled Mangio, sorry
But what its Python?
low shard
Python is the programming language that rvc uses, download it by there https://www.python.org/downloads/
pallid patio
Ookay, what I do now? Just run the applio?
rare gobletBOT
Ayo? @pallid patio level 4 !!!
pallid patio
Traceback (most recent call last):
File "C:\Users\name\Downloads\Applio-main\Applio-main\app.py", line 1, in <module>
import gradio as gr
ModuleNotFoundError: No module named 'gradio'
dawn nebula
Traceback (most recent call last):
File "C:\Users\name\Downloads\Applio-main\A...
where did u download it
dawn nebula
https://docs.aihub.wtf/essentials/whats-rvc/#forks
i think its better to use the prezip
pallid patio
i think its better to use the prezip
Whats the difference? I'm still learning ;-;
dawn nebula
the traditional applio install method can have many things go wrong while nomal users cant fix it. I would prefer a prezip since its more stable
pallid patio
the traditional applio install method can have many things go wrong while nomal ...
Oh, makes sense
I just need Python to run it? Or smtg more?
pallid patio
I've dowloaded it
dawn nebula
extract
pallid patio
Ookay, done
stable crest
How do you train a ai rvc model without an nvidia gpu?
rotund prawn
or Rigel v0.1 (it's not exist but🤦♂️ )
violet heron
How do you train a ai rvc model without an nvidia gpu?
Google colab
stable crest
Google colab
Google colab link?
low shard
Google colab link?
Check our docs https://docs.aihub.wtf
Google colab platform offers u 4 hours daily of free T4 Gpu
It uses a remote google pc basically
violet heron
Google colab link?
-colab RVC Disconnected
azure marshBOT
-colab RVC Disconnected
Suggestions for @stable crest
violet heron
Guides in #1159513888199540817 or https://docs.aihub.wtf/
fleet sleet
Check our docs https://docs.aihub.wtf
Can you tell me if there is any pretrained in Portuguese( take a voice model in English and make it speak Portuguese fluently)?
low shard
Can you tell me if there is any pretrained in Portuguese( take a voice model in ...
as of now and of what i know, there isn't any portuguese pretrains, but if you want you could take a look at #1233407331405004954 and see if there is any pretrains that could be maybe be phonetically close to the portoguese language
fleet sleet
Pretrain
fleet sleet
as of now and of what i know, there isn't any portuguese pretrains, but if you w...
How can I make a pretrain?
Is it hard?
low shard
How can I make a pretrain?
scroll the #1233407331405004954 guide, its said how to make them
low shard
Is it hard?
you need very good gpu + tons & tons of hours for the dataset, the rest is basically the same of making a model
fleet sleet
you need very good gpu + tons & tons of hours for the dataset, the rest is basic...
Ok, thank you
low shard
yw
odd shale
Ok, thank you
You need also a nice variety of speakers when making the dataset for your pretrain
Also, make sure to balance between clean audio and noisy audio.
Because if you use only clean audio, your pretrain will be pretty sensible to noisy datasets
fleet sleet
Because if you use only clean audio, your pretrain will be pretty sensible to no...
Thank you for the information
rotund prawn
You need also a nice variety of speakers when making the dataset for your pretra...
Hi
Do I need to use SnowieV3.1 pretrain for Japanese dataset?
odd shale
Do I need to use SnowieV3.1 pretrain for Japanese dataset?
Ig you can use it if you want.
rotund prawn
Okey
Thanks
fleet sleet
You need also a nice variety of speakers when making the dataset for your pretra...
I made a model in Portuguese, it was an English speaking person who I first cloned his voice normally and then I made inferences with audio from Elevenlabs in Portuguese and a song, I have as goal to remove the American accent(English accent) but forget that one audio file was in mp3 and the new model came out bad in quality and it still has the accent, do you think that the problem was the mp3 audio file?
rare gobletBOT
Ayo? @fleet sleet level 4 !!!
fleet sleet
I created first a model(English) then I tried to made a fluent portuguese model out of the first one(inference audios).
pliant cobalt
cell 1 loading forever how can i fix it
fleet sleet
In the graph it showed that the line was in the top not going down
odd shale
cell 1 loading forever how can i fix it
Which colab are you using?
pliant cobalt
Which colab are you using?
rare gobletBOT
Ayo? @pliant cobalt level 1 !!!
burnt zealot
which one do i use
odd shale
https://colab.research.google.com/github/hinabl/voice-changer-colab/blob/master/...
Yeah.. I'll ask Shad.
@pastel oak
Use De-Echo-DeReverb model in UVR or MVSEP
pastel oak
cell 1 loading forever how can i fix it
It takes longer depending on your internet speed.
- Make sure your internet is not disconnecting randomly
- Make sure if it asks you to access something like google drive, you say yes and dont click away from it
If it still takes a long time, try a different browser
odd shale
Sorry, there's no other dereverb methods.
Unless you use dereverb module in RX10 but that program is expensive and the module itself isn't really suitable for music.
You're not going to find other way.
pliant cobalt
It takes longer depending on your internet speed.
- Make sure your internet is ...
my internet is fine it was working and after i refreshed this started
calm torrent
my voice ai wont take the input in discord help me
odd shale
my voice ai wont take the input in discord help me
We don't give support for voice.ai, sorry.
Sorry ig.
I've already suggested you to use De Echo De Reverb model in MVSEP.
Yeah, its the same model as UVR one
calm torrent
We don't give support for voice.ai, sorry.
so wht application
severe violet
rotund prawn
Model maker?
glass drum
is using batch size 4 a bad idea for a 20 mins dataset?
proper shale
is using batch size 4 a bad idea for a 20 mins dataset?
might be slightly unstable, but sure, try that
proud elbow
rtx 3070 might not have enough vram
it is enough for batch size up to 8
rare gobletBOT
Ayo? @proud elbow level 35 !!!
proper shale
rtx 3070 might not have enough vram
nah it does
glass drum
it is enough for batch size up to 8
oh I thought lower means more vram. guess lower just means slower processing
proper shale
it is enough for batch size up to 8
probably can handle 12, ive heard some people saying that w 8GB VRAM cards
proper shale
oh I thought lower means more vram. guess lower just means slower processing
lower = less samples trained at once
glass drum
I see, if the accuracy is better sounds worth it
proper shale
I see, if the accuracy is better sounds worth it
It kinda is, but it also makes training kinda unstable... so be careful
or find a good batch that doesn't make ur training so unstable
glass drum
It kinda is, but it also makes training kinda unstable... so be careful
or find...
what does unstable mean? like how can I detect it
using applio btw
proper shale
basically, smaller batch means the model parameters change more per training cycle (epoch) than higher batches
this also causes instabilities if too low
proper shale
yeah I use that
check g/total, see if it's spiky at 0.9 smoothing
glass drum
alright 
proud elbow
what does unstable mean? like how can I detect it
lower batch is usually more suitable to shorter datasets
proper shale
lower batch is usually more suitable to shorter datasets
this tends to vary tho...
like even big datasets can like low batches sometimes
every model's different
proud elbow
this tends to vary tho...
yea it can vary between 4, 8, 16
proud elbow
this tends to vary tho...
though I need some technical explanations on if non-power of 2 numbers are also fine actually
proper shale
though I need some technical explanations on if non-power of 2 numbers are also ...
yeah
AI works better with those but eh it also works with others too
glass drum
so far so good. guess if it stays stable i'll wait around 6k~ steps and try the model. if bad will try batch = 8
proper shale
so far so good. guess if it stays stable i'll wait around 6k~ steps and try the ...
oh yeah seems pretty... stable
glass drum
noice 
wraith coral
is there a tutorial for rvc
calm jackal
-rvc
rare gobletBOT
Ayo? @calm jackal level 3 !!!
calm jackal
-colab
azure marshBOT
RVC Guides (How to Make AI Cover)
rotund prawn
oh yeah seems pretty... stable
Hi
static forum
-colab
proper shale
that's probably the filtering's fault
foggy folio
what is the best way to make covers now?
proper shale
what is the best way to make covers now?
via colab? using applio or ilaria rvc ig
foggy folio
hmm is ilaria or applio better
proper shale
you know I think huggingface is better tbh ngl
huggingface uses cpu unless u pay for a gpu
mortal beacon
Is there somewhere to download the base RVC - V2 model? I see the Applio, Mainline, and Mangio models, but aren't these V1?
violet heron
Is there somewhere to download the base RVC - V2 model? I see the Applio, Mainli...
All those have v1 and v2 support
mystic marsh
-colab
mortal beacon
Pardon my ignorance, but is collab calling a model via an API? What is the difference between the "official" rvc-2, RVC-HFv2, and RVC v2 Huggingface version?
I'm looking for the base model to download locally, rather than train a model through an API and only have the outputted singular trained model
snow rose
@graceful obsidian accept my friend request beacuse i cant message you beacuse you turned this mode where non friends cant message you i think
foggy folio
why is applio not opening locally?
graceful obsidian
@snow rose ps.
Sometimes reading or knowing your own account's settings doesn't hurt:
graceful obsidian
" accept my friend request beacuse i cant message you beacuse you turned this mode where non friends cant message you i think "
It is not that I have it on, it is on your side.
aka, you're the one having it disabled, not me.
best of luck ~ 👋
snow rose
" accept my friend request beacuse i cant message you beacuse you turned this mo...
i can message everyone
graceful obsidian
¯_(ツ)_/¯
snow rose
but not you
graceful obsidian
🤦♂️ Wouldn't be so sure
literally every single safety " lock " on my acc is off
It'd be appreciated if you read a lil more on how discord works for future
graceful obsidian
either way, cheers ~
cya
snow rose
i love how you are so rude on dms but so kind on server
graceful obsidian
It ain't what you call rude, it is called being direct af and respecting your own time.
What's rude is me offering you help which I don't really do nowadays anymore, and you leaving without any response or callback for days alr
Aside, tf you expect when you " xd " on me - if there's one thing you gotta know 'bout me, I fucking hate that sort of people.
Once again, best of luck ~ cheers ~ cya.
snow rose
It ain't what you call rude, it is called being direct af and respecting your ow...
you started this
then i responded if he is mad that i messaged him when he is doing something
he started insulting me
like wtf
graceful obsidian
- First of all, don't start dramas on a help channel.
- You called me a " materialist " just because I simply sent you a link to my comms page since you wanted a free model despite me NOWHERE offering free models. That was lame and fully disrespectful towards my workflow and commissions. I am a person that's doing paid commissions and not free models, calling me a materialist mainly due to the fact I expect payment for a job I typically in a time span ranging from 2 days to 2 weeks seems fair, valid and up to my standards and a belief where I value my time over anything else in social contacts, especially if it's about a random like you that brings nothing to my life.
- You added me despite me removing you - this should be a clear indicator to you I've no interest in your presence. It's not disrespecting my private space.
- I offered you my help, you more or less ignored it. This ain't my issue. You mentioned you found a guy that's gonna make you a model or help with music? ( don't remember ) that's cool, but it ain't my business.
- I am a very darn direct person and just because you're new here ( I've been a part of AI hub community even before any forks started, when only mainline and mangio existed ( and was still maintained by mainly mangio ). What I am coming from, you're not allowed to call me unkind without even fucking knowing me or publicly disrespecting me.
- You literally told me to " fuck off " after I asked if you have anything to add before being unfriended ( that's what my toxic ex would do lol - as I stated in the dm )
- I responded to your ping explaining my discord settings allow for any kind of DM or users interaction and that the issue is on your side.
- Once again, if you've no chat history properly translated or any valid facts to back up your any bs statements, stay away from it.
Almost forgot. Learn to say " thank you " when a person purely out of a nice kind, takes 2 or was it 3 hours? into trying to fix a cover mess you've made. ✨ cheers.
( and ye, if you have any personal issues, take it to the fucking dm. Seen too many attention or drama queens in my life. )
snow rose
you are funny
graceful obsidian
Literally, rot brain case.
Step away or Imma ping moderation.
once again, this is a help channel take your emotive issues to dm.
@proper shale Soz for disturbing if anything
can you have a look in here?
I'd rather avoid any unnecessary drama
snow rose
Step away or Imma ping moderation.
***once again, this is a `help channel`*** ...
i can send them our whole dm if you dont stop talking about me so yeah you can ping them
graceful obsidian
Don't worry, I can literally translate everything and send to the moderation.
We can even hop into vc :cheers: ✨
Am tired af of toxic polish girls not understanding I value my time and my job is not free, it is called commissions.
snow rose
Don't worry, I can literally translate everything and send to the moderation.
We...
idc about your comissions, stop flexing with them XD
graceful obsidian
Mj, can you, please?
graceful obsidian
The girly literally cannot read
3rd time or so saying this is a help and not drama channel, smh
proper shale
yeah guys, take it to dm's, please
graceful obsidian
^ thank you. Hopefully now she'll listen.
snow rose
yeah guys, take it to dm's, please
he blocked me (i would do same if he didnt)
graceful obsidian
XD?
you do realize that when you have settings set as you do and you're not friends with someone, the system notifications of discord act as if you were blocked right? and I am 100% sure most of users here can confirm that.
Now, if that's all you need. step back and if you have any frustrations, take it to the dm as MJ mentioned.
snow rose
you do realize that when you have settings set as you do and you're not friends ...
you still continue this for some unknown reason
graceful obsidian
sigh @proper shale
and you, type some. Learn to differentiate current time:
that was, for christ sake, 7 minutes ago lmao.
ps, emotive issues = " An emotive situation or issue is likely to make people feel strong emotions. "
proper shale
mighty parcel
rotund cairn
Why does all youtubers when they use okada voice changer it sounds so good and when i use it, its not that good
proper shale
Why does all youtubers when they use okada voice changer it sounds so good and w...
it's the model or the settings you're using
also, mic quality 
graceful obsidian
I'd take it to #🔍│help-w-okada
but I can tell you right away, it's 3 potential things:
- Model's quality
- proper setting for chunk and all
- impersonation sometimes helps the model accomodate for impurities or imperfections
@rotund cairn
but ye, Mj def knows more about w-okada than me so
rotund cairn
I'd take it to <#1159290161683767298>
but I can tell you right away, it's 3 pot...
i can send picture of my settings
proper shale
yeah, you can
violet heron
please do
graceful obsidian
oh, btw, the tune thing
that's essentially semitones changing for your model's output
so if your pitch / formant distribution is different than model drastically
you gotta play with it
if your voice is lower than model's, rise the " tune " param
if it's higher, lower the tune
proper shale
^
graceful obsidian
also, noise suppression can degrade the stability by a lot
try without em ( sup1 and sup2 )
proper shale
ehm sup2 might be essential depending on how noisy ur place is
graceful obsidian
@rotund cairn also, check this:
those are proposed values per-gpu
try: 112 chunk +16384 extra
if it lags: 128 chunk +16384 extra
proper shale
<@1037411435455713340> also, check this:
id also say to experiment with extra, since it can make results better
👀
graceful obsidian
yeee
extra is essentially extra ( in memory, kinda overhead ) inference
because realtime voice changing is in fact, just, live inferencing on your voice
vs static input such as .wav or whatever ( rvc )
ofc, it increases the delay but might help for weaker gpus / cases that can't handle the chunk
proper shale
extra is essentially extra ( in memory, kinda overhead ) inference
yeah i believe it's something like that
extra time to process the thing
graceful obsidian
and ( as much as realtime infer gui differs )
chunk must be how many chunks you feed the w-okada
to infer + fade / overlay
so if it's powerful enough and quick, it can take in more chunks ( smaller chunks so, smaller values of length / duration ) efficiently
if it's not powerful enough, it won't make the " live inferencing " in time, resulting in choppiness
so I'd say, perhaps picking either not too crazy chunk is a good ideal 128 - 2xx ? and giving it 0.5 to 1.5s of extra infer ( extra value )
rotund cairn
alr now i have 2 question what should i change in the in/out settings and where can i find a really good voice
proper shale
alr now i have 2 question what should i change in the in/out settings and where ...
in/out are for volume basically
and you can always check #1175430844685484042 and the samples they give there
or weights.gg
ofc make sure you're actually seeing RVC models, gpt-sovits models wont work
graceful obsidian
sorta like that
rather simply shown
smaller they are = more of em
more of em = higher gpu's performance required to keep up
but then, if gpu can keep up ( and provided you don't rely too much on extra param ) = smaller the delay and potentially smoothed out pitch ( f0 ) integrity
or so I interpret it at least, someone's more into w-okada or rvc's native realtime might add something in or correct me ~
graceful obsidian
alr now i have 2 question what should i change in the in/out settings and where ...
ye as mj said, additionally.. well, given how weights.gg handles " demo samples ", it's really a hit or miss
when it comes to finding hq models
they're just not per-case finetuned, which doesn't show the model's performance in " optimal " scenario ( which could be rvc inference or
real time voice changing (( provided it's set correctly in the w-okada itself )) )
flint seal
is there anyway to clear the model list on ilaria rvc? the models just keep piling up and its rlly annoying
flint seal
ohhhh
graceful obsidian
hope that works
proper shale
is there anyway to clear the model list on ilaria rvc? the models just keep pili...
this is a public space, so thats why theres models there
so duplicate it to make it private
flint seal
this is a public space, so thats why theres models there
oh alr thanks
proper shale
oh alr thanks
yw :)
flint seal
it WORKED
FINALLY
been having this issue for months
cant believe ive never thought of tha
graceful obsidian
@barren acorn ur gpu?
barren acorn
gtx1660
graceful obsidian
what's your chunk and extra values?
barren acorn
192
barren acorn
yes
graceful obsidian
hmmm...
barren acorn
i put them like that
graceful obsidian
can you screenshot the ui?
barren acorn
what ui?
graceful obsidian
u use w-okada? or rvc's native voice changer
barren acorn
yea
graceful obsidian
so, w-okada?
graceful obsidian
you got the right release of w-okada?
onnx cuda
barren acorn
yea they send it to me here
barren acorn
hmmm.. same for models?
every model
barren acorn
converted pytorch to onnx?
yes
graceful obsidian
alright, so then, please ss for me your w-okada's settings
I use rvc's native so don't have w-okada interface memorized
barren acorn
graceful obsidian
Hmmmm.. try to use a chunk value that gives around 1 second of delay
1024ms most likely, will be written
barren acorn
Alr
barren acorn
just debug if it'll be cutting as well
How
barren acorn
It stopped
graceful obsidian
as in, isn't cutting / lagging anymore?
pastel oak
https://huggingface.co/wok000/vcclient000/resolve/main/MMVCServerSIO_win_onnxgpu...
this is outdated btw
dont follow youtube tutorials
stiff goblet
training should be 40k sample rate right?
pastel oak
oh
f0 det rmvpe
chunk and extra are fine, can probably use 128 chunk
output CABLE (download vac btw cable is dogshit)
tough fiber
packages/ffmpeg/_run.py", line 325, in run
raise Error('ffmpeg', out, err)
ffmpeg._run.Error: ffmpeg error (see stderr output for detail)
how can i fix this
start preprocess
['trainset_preprocess_pipeline_print.py', '/notebooks/ixtivhdata', '32000', '8', '/notebooks/Mangio-RVC-Fork/logs/cybprin', 'False']
/notebooks/ixtivhdata/.ipynb_checkpoints->Traceback (most recent call last):
File "/notebooks/Mangio-RVC-Fork/my_utils.py", line 134, in load_audio
ffmpeg.input(file, threads=0)
AttributeError: module 'ffmpeg' has no attribute 'input'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/notebooks/Mangio-RVC-Fork/trainset_preprocess_pipeline_print.py", line 84, in pipeline
audio = load_audio(path, self.sr, DoFormant, Quefrency, Timbre)
File "/notebooks/Mangio-RVC-Fork/my_utils.py", line 141, in load_audio
raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: module 'ffmpeg' has no attribute 'input'
/notebooks/ixtivhdata/fffff.wav->Traceback (most recent call last):
File "/notebooks/Mangio-RVC-Fork/my_utils.py", line 134, in load_audio
ffmpeg.input(file, threads=0)
AttributeError: module 'ffmpeg' has no attribute 'input'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/notebooks/Mangio-RVC-Fork/trainset_preprocess_pipeline_print.py", line 84, in pipeline
audio = load_audio(path, self.sr, DoFormant, Quefrency, Timbre)
File "/notebooks/Mangio-RVC-Fork/my_utils.py", line 141, in load_audio
raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: module 'ffmpeg' has no attribute 'input'
end preprocess
jade cove
<@&1202347671260762142>
glossy grove
Is there a good guide for multi speaker rvc using gradio/applio?
azure marshBOT
RVC Guides (How to Make AI Cover)
glossy grove
Yeah that's the basic bit. Doesn't really tell me what I need sadly
odd shale
Is there a good guide for multi speaker rvc using gradio/applio?
multispeaker?
quaint tendon
just wondering if I have 15 minutes of data size, what batch size should I use?
low shard
just wondering if I have 15 minutes of data size, what batch size should I use?
I think 8 should be good considering the datasset varies
quaint tendon
I think 8 should be good considering the datasset varies
alright! thank you so much!
low shard
Yw
glossy grove
multispeaker?
Yes you can trian an rvc model with data from more that one speaker , or person. I just don't know how in applio
wanton root
Guys hello, can anybody can help me?
i have problem like this
Pipeline is not initialized
Waiting generate pipeline
glossy grove
I'll just poke shoes tonight if he's on
torpid steppe
I need help, it won't let me use the RVC voices, the Beatrice ones do but the rest don't
upbeat coral
is it better to sing or speak if you are trying to train your own voice to become a voice model for song covers?
simple mango
Anyone have an idea of why the illaria rvc just errors no matter what file I do like it was just working and now it errors every day for like a week straight
upbeat coral
Anyone have an idea of why the illaria rvc just errors no matter what file I do ...
I get errors too. it doesn't work
no one seems to know the answer either
I am just waiting for the next fork to come out
simple mango
dang that really sucsk
rapid niche
yeah public url isn't working for me either
azure marshBOT
clear snow
googled the issue but mangio wont give me my PTH file for some reason. Yes the batch is on 1. No i cant look for it because the model i have dosent work with Mangio or normal GUI. They said turn batch down well i cant do that anymore so clearly theres a problem im too small brained to understand. And of course its the first model i've had to make in awhile and i cant read the dumb code
problem is it dosent look like its training.......despite finishing the other processes
yes the index file exists but my code isnt giving me numbers
i dont think 'TypeError: init() takes 19 positional arguments but 20 were given
write filelist done' counts as numbers
rare gobletBOT
Ayo? @clear snow level 1 !!!
clear snow
starting to think the chinese/japanese/some other language text means 'theres an issue we wont tell you :3'
if somebody on these forums says 'same issue' im gonna crawl through their screen and put snakes in their pillows
acoustic compass
if somebody on these forums says 'same issue' im gonna crawl through their scree...
same issue
clear snow
stop
ive about had it and im serious
the one time i have to make something myself my computer is laughing at me in 'No NvIdIa GpU' WHICH IVE SPECIFIED I DONT HAVE,
im pretty sure i cant put up images
so
like
Nvm but look. Tis empty. like my garden of [no]'s to give
the writing of the filelist. and no my Weights folder hasnt changed
rare gobletBOT
Ayo? @clear snow level 2 !!!
clear snow
arclame for the love of GOD i need help not a level up notificiation
trying to use the moddel i found results in this.
violet heron
trying to use the moddel i found results in this.
RVC-GUI is very outdated
Use original RVC or a fork like applio instead
-local
azure marshBOT
violet heron
Wait
clear snow
Tried to use Mangio. as seen above its making me blow up
violet heron
Tried to use Mangio. as seen above its making me blow up
Did you try mainline?
clear snow
Have no idea what that is
violet heron
Have no idea what that is
odd shale
Have no idea what that is
RVC Mainline.
odd shale
https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/releases/t...
That one.
junior halo
Someone remind me how the #1159514067187277865 channel works?
clear snow
https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/releases/t...
it only shows realtime in the files and i cant find the bat file to run the normal web version. either im dumb or im dumber
foggy perch
Is there a preferred version to use (original RVC vs Mangio vs Applio, etc)?
distant marten
wispy burrow
Someone remind me how the <#1159514067187277865> channel works?
You submit your Model there.
If it passes the Quality Control, one is granted the Model Maker Role.
That Role allows you to share your Models in #1175430844685484042 .
wispy burrow
Is there a preferred version to use (original RVC vs Mangio vs Applio, etc)?
Not really.
You can use whatever you want.
junior halo
How's my training looking so far? (Training to 500 epochs. 8 steps. 8 minutes of data.)
azure marshBOT
RVC Guides (How to Make AI Cover)
royal jacinth
does anyone know how to enable public url on applio linux
tough fiber
start preprocess
['trainset_preprocess_pipeline_print.py', '/notebooks/ixtivhdata', '32000', '8', '/notebooks/Mangio-RVC-Fork/logs/cybprin', 'False']
/notebooks/ixtivhdata/.ipynb_checkpoints->Traceback (most recent call last):
File "/notebooks/Mangio-RVC-Fork/my_utils.py", line 134, in load_audio
ffmpeg.input(file, threads=0)
AttributeError: module 'ffmpeg' has no attribute 'input'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/notebooks/Mangio-RVC-Fork/trainset_preprocess_pipeline_print.py", line 84, in pipeline
audio = load_audio(path, self.sr, DoFormant, Quefrency, Timbre)
File "/notebooks/Mangio-RVC-Fork/my_utils.py", line 141, in load_audio
raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: module 'ffmpeg' has no attribute 'input'
/notebooks/ixtivhdata/fffff.wav->Traceback (most recent call last):
File "/notebooks/Mangio-RVC-Fork/my_utils.py", line 134, in load_audio
ffmpeg.input(file, threads=0)
AttributeError: module 'ffmpeg' has no attribute 'input'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/notebooks/Mangio-RVC-Fork/trainset_preprocess_pipeline_print.py", line 84, in pipeline
audio = load_audio(path, self.sr, DoFormant, Quefrency, Timbre)
File "/notebooks/Mangio-RVC-Fork/my_utils.py", line 141, in load_audio
raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: module 'ffmpeg' has no attribute 'input'
end preprocess
how can i fix this
proud elbow
start preprocess
['trainset_preprocess_pipeline_print.py', '/notebooks/ixtivhdat...
having one or several wav files in a folder as input dataset shouldn't throw errors like this, though I have no much idea since I stopped using mangio and use the latest applio instead
tough fiber
having one or several wav files in a folder as input dataset shouldn't throw err...
poff
.paperspace
what was the command
tough fiber
bro it reached 40 epoch in 2 mişn
applio is sh**t
whats the problem idk mangio is better
pastel oak
bro it reached 40 epoch in 2 mişn
if your dataset is very short and batch size very high, this is realistic
How are you concluding its shit based on numbers
tough fiber
How are you concluding its shit based on numbers
it looks do not train
even this tensorboard thing on applio
what is this really
i cant open :D
mangio old but works pretty fine
@pastel oak so is it normal for batch size 16. and my data set 1 hour 20min
it reached 40 epoch in 2min
pastel oak
<@965636299509882980> so is it normal for batch size 16. and my data set 1 hour ...
ah nah then its not normal
tough fiber
ah nah then its not normal
yea i dont know wheres problem
whatever shouldnt i install applio i did mistake lol
ill install mangio again
proud elbow
bro it reached 40 epoch in 2 mişn
is it colab or local, or hosted in somewhere?
the local applio is pretty fine on me, even have tried the new TITAN pretrain
tough fiber
i pasted my data path folder so idk it isnt work
pastel oak
<@965636299509882980> so is it normal for batch size 16. and my data set 1 hour ...
@odd shale You use Paperspace right?
Person has a 1:20h dataset, 16 batch size, and each epoch takes 2 seconds. Is that normal?
odd shale
<@628603171735994379> You use Paperspace right?
Person has a 1:20h dataset, 16 ...
I'm not sure if its the gpu he used or the batch size he selected.
I need to see his settings.
Batch 8 is normally enough.
proper shale
Filtering of the vocals. They definitely had reverb and that's why they got those weird artifacts in the beginning
azure marshBOT
RVC Guides (How to Make AI Cover)
verbal quiver
Hi everyone,
I have a beloved idol who recently faced online censorship. My wife and I are getting married this year, and it was through our shared admiration for this idol that we first connected. I want to create our love song using an AI voice model, but after researching, it seems that my GPU (3050) is insufficient for the task according to common websites. I hope to find assistance from experts here or someone who can help me with the production. Your help would be greatly appreciated.
Thank you.
ember lily
Someone have problem with ilaria rvc? There is only the local url and not the public :/
@brave garnet
verbal quiver
rare gobletBOT
Ayo? @verbal quiver level 1 !!!
verbal quiver
proper shale
Someone have problem with ilaria rvc? There is only the local url and not the pu...
try to use https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb
proper shale
Hi everyone,<:e_boorgir:1159654275069255750>
I have a beloved idol who recentl...
If there's a model of said idol somewhere, you could just infer (convert vocals) locally.
Or you could train models via Colab:
https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS
Check https://docs.aihub.wtf/ for guides on them too ;)
pliant tulip
why is applio taking forever uploading the index/pth and audio?
gaunt trout
Sorry if this is the wrong channel to ask this but has anyone been having issues getting the collabs to work? I was using it to make an ai cover and it got up to 5.8k seconds before I just gave up
pliant tulip
ilaria collab seems to be down right now
applio is the only other rvc v2 one active rn but not working for me rn
abstract kestrel
Could not create share link. Please check your Internet connection or our status page: https://status.gradio.age..
Also please ensure that your antivirus or firewall is not blocking the binary file located at: /usr/local/lib/python3.10/dist-packages/gradio/froc_linux_amd64_v0.2. Help me!!!
pastel oak
Could not create share link. Please check your Internet connection or our status...
It would be beneficial if you explained what application you are running, when this happens, anything
abstract kestrel
It would be beneficial if you explained what application you are running, when t...
I don’t understand
abstract kestrel
Colab
abstract kestrel
Ilaria
pastel oak
Ilaria is down right now doesnt work
abstract kestrel
Mainline ver
pastel oak
abstract kestrel
https://docs.aihub.wtf/rvc/local/mainline/
pastel oak
Ilaria is down right now doesnt work
Ilaria is down right now doesnt work
Ilaria is down right now doesnt work
Ilaria is down right now doesnt work
abstract kestrel
Ilaria is down right now doesnt work
Ilaria is down right now doesnt work
Ilaria...
Tell me about other colabs
Tell me about other colabs
Tell me about other colabs
pastel oak
abstract kestrel
https://docs.aihub.wtf/rvc/cloud/applio-colab/
Hahahahahahahaha
WTF!!!!!!!
What is up password?
open mountain
-colab
pastel oak
Hahahahahahahaha
it works for me
as Light said, you had to wait for a few more seconds and it shared the link + passwort ip
lusty leaf
hello guys. I am creating rvc model, but writing: Data preprocessing error... Make sure that the folder with the dataset is selected correctly
i make model in https://colab.research.google.com/drive/1030ibBABNaNZ3CjIoDXstcCR7VhEnSAA#scrollTo=c9a4PKyP1yQE
Can you are help me?
pastel oak
Hahahahahahahaha
open share link, enter password ip
lusty leaf
hello guys. I am creating rvc model, but writing: Data preprocessing error... Ma...
but i downloading in dataset wav file
can you are help me?
pastel oak
I dont know that colab you linked sorry
lusty leaf
I dont know that colab you linked sorry
okay(
lusty leaf
I dont know that colab you linked sorry
can you say, collab for create ai model?
pastel oak
can you say, collab for create ai model?
-colab
pastel oak
RVC Disconnected or Applio works
lusty leaf
thnx
red kayak
Hi everyone,<:e_boorgir:1159654275069255750>
I have a beloved idol who recentl...
nope u can run RVC on a 3050 just fine
proper shale
nope u can run RVC on a 3050 just fine
training will be 
red kayak
training will be <:distress:1159577413244686407>
but still it can train
ppl can even train on a gtx 1080
proper shale
but still it can train
2 batch will be hell and will prolly lag the system yk
proper shale
ppl can even train on a gtx 1080
has more VRAM so
...ik its slow as balls but not as much as 3050 just because of the extra vram
red kayak
2 batch will be hell and will prolly lag the system yk
assuming they have the ram....
proper shale
assuming they have the ram....
most 3050s are 4GB, if they are lucky and have the 8GB version eh maybe